In January 2001, the Basel Committee issued CP-2, which set forth the promised IRB approach (Basel Committee 2001a, 2001b). In fact, it set forth two IRB methodologies, reflecting the committee’s new view that small and medium-sized institutions should have the option, if they were willing and able to invest the necessary resources, to elect such an ap- proach. The committee indicated that it would “finalize” the new accord by the end of 2001, for implementation by 2004. As it turned out, the re- lease of CP-2 was only the first step down what has been an arduous path to the committee’s goal of implementing an IRB regulatory paradigm in its member countries.
CP-2 changed the nature of the Basel II exercise. In place of the ap- proximately 60-page CP-1 and the 30-page Basel I was a package of over 450 pages, including the proposal itself and seven supporting documents.
The proposal was not just long and detailed but technically complex, characteristics that have carried over into the final version of Basel II. De- spite its length and complexity, however, CP-2 was fundamentally in- complete. Several key issues were left for later elaboration.35Furthermore, the committee acknowledged that it was uncertain about the proper cali- bration of the IRB risk-weighting formulas. Accordingly, it did “not have enough information at the moment to assess the full impact of the pro- posal” (Basel Committee 2001a, paragraph 52). As would later become clear during a series of quantitative impact studies (QIS), the committee
34. See William McDonough, “Update on the Major Initiative to Revise the 1998 Capital Ac- cord,” remarks to the 4th Annual Supervision Conference of the British Bankers Association, June 19, 2000.
35. Indeed, the committee characterized its approach as “evolutionary,” anticipating revi- sions of its rules as risk management techniques developed further. Thus, even with respect to matters it did address in CP-2, the committee raised the prospect of regular changes.
has never really solved the problem of predicting the impact of its pro- posals on required capital levels.
Thus, despite its length, CP-2 was less a fully articulated plan resem- bling a proposed regulation in US administrative law practice than, in the committee’s words, “a starting point for additional dialogue” (Basel Com- mittee 2001a, 53). While the committee used this language to refer only to the issue of calibrating the risk weights, it may as well have referred to the entire proposal. The gaps in CP-2 and numerous elements of the proposal that were included elicited another round of strong criticism—principally from banks but also from consultants and commentators. These com- plaints led to abrupt shifts of position by the committee on some fairly ba- sic issues and a series of delays in completing the revised framework.
The delays and extensive revisions not only called into question the wisdom of the committee’s approach but also gave each bank or group of banks in the member countries ample time to organize in pursuit of changes in the proposal that would be to their liking. The result was a se- ries of ad hoc accommodations by the committee to a wide variety of de- mands for change. By releasing CP-1 and CP-2—two fundamentally in- complete proposals—the Basel Committee ensured that it would spend much of the Basel II process reacting to criticism and on the defensive. In a sense, the committee became captive to its own process, forging in- domitably ahead in order to avoid the failure of an abandoned project but needing constantly to placate the banks’ opposition if it was to make progress.
Although a proposal as detailed as CP-2 defies easy summary, its key elements are set forth in box 4.2. Notwithstanding the extensive dispar- agement of the CP-1 proposal, the basics of the “standardized” approach to capital requirements remained the same, though several modifications were made, most for the stated end of increasing risk sensitivity. Among the more important were the addition of a fourth 50 percent risk bucket for corporate exposures, a preferential risk weight for short-term expo- sures to other banks denominated in the local currency, and the possibil- ity that a bank or corporation could have a lower risk weighting than the sovereign in its country of incorporation. In response to the criticism that lending to developing country sovereigns would be seriously affected by the absence of external credit ratings for most such countries, the commit- tee specified that export credit agencies were a permissible source of credit scores for sovereign exposures.
The committee’s key innovations in CP-2 were the IRB proposals, which envisioned two distinct approaches. Creation of the “foundational”
internal ratings–based (F-IRB) approach responded to the wider-than- expected demand among banks in member countries for participation in an IRB approach. The A-IRB approach was assumed to be applicable only to a relative handful of big banks. Thus was established the second key structural feature of Basel II. The first was the three-pillar approach
Box 4.2 Changes proposed in the second consultative paper
Pillar 1: Minimum Capital Requirements Standardized Approach
䡲 addition of export credit agencies as source of ratings for sovereign risk 䡲 adjustments to risk weight proposed in the first consultative paper (CP-1) (e.g.,
exposure to a bank may be rated as less risky than that bank’s sovereign) 䡲 wider range of collateral acceptable for credit risk mitigation
Internal Ratings–Based (IRB) Approach 䡲 creation of two IRB approaches
䡲 foundational approach uses banks’ estimates of probability of default and su- pervisory parameters for exposure in the event of default, loss if default occurred, and maturity of the exposure in calculation of risk weights for each sovereign, bank, and corporate exposure
䡲 advanced approach permits use of bank estimates of probability of default, exposure in the event of default, loss if default occurred, and maturity of the exposure
䡲 capital requirements determined through supervisory formulas into which probability of default, exposure in the event of default, loss if default occurred, and maturity of the exposure are inputs
䡲 separate treatment of retail exposures based on risk associated with distinct segments of retail lending
Asset Securitization
䡲 both standardized and IRB approaches 䡲 clarification of “clean break” requirement
䡲 differentiated capital requirements for originating, investing, and sponsoring banks
Operational Risk
䡲 basic indicator approach to calculate requirement using percentage of a single indicator (e.g., gross income)
䡲 standardized approach to use different indicators and percentages for differ- ent business lines
(box continues on next page)
adopted in CP-1. Now CP-2 had introduced three separate methodolo- gies for banks of varying risk management capabilities. This feature ap- plied not just to the calculation of risk weights but also to other key ele- ments of pillar 1.
The essence of the IRB approaches was that qualifying banks could use their own estimates of the probability of default of each of its expo- sures. Under the A-IRB approach, banks would also be able to use their own estimates of their exposure in the event of default, loss if default oc- curred, and maturity of the exposure. These values would be converted into risk weights through formulas devised by the supervisors. Separate formulas were provided for corporate, sovereign, and bank exposures. A somewhat different framework would apply to retail exposures.36The com- mittee indicated that separate frameworks were also necessary for project finance and equity exposures but that it had not yet developed them.
Box 4.2 Changes proposed in the second consultative paper (continued)
䡲 internal measurement approach to use combination of internally-developed data and supervisory factors
Pillar 2: Supervisory Review of Capital Adequacy 䡲 elaboration of four principles from CP-1
䡲 interest rate risk to be evaluated as part of supervisory review rather than subject to capital charge calculation
Pillar 3: Market Discipline
䡲 elaboration of required disclosures for:
⽧ scope of application;
⽧ composition of capital;
⽧ risk exposure assessment and management processes; and
⽧ capital adequacy.
䡲 distinctions between “core” and “supplementary” disclosures 䡲 expectation of supervisory response to inadequate bank disclosures
Source:Basel Committee (2001b).
36. Since banks do not customarily assign ratings to individual retail exposures, the com- mittee chose to follow bank practice in allowing the grouping of exposures into categories, or “segments,” based on similar risk characteristics.
To be eligible for an IRB approach, a bank would have to meet mini- mum requirements relating to its internal rating, credit assessment, and disclosure practices. To qualify for the A-IRB approach, a bank would have to meet additional requirements applicable to the calculation of ex- posure in the event of default, loss if default occurred, and maturity of the exposure. In proposing reduced capital requirements for credit risk miti- gation techniques such as collateral and credit derivatives, the committee also set forth different methods, with the A-IRB approach again allowing the greatest use of internal assessments.
The committee provided special rules for calculating capital require- ments for securitization exposures. Reflecting the concern that securitiza- tion was a form of regulatory arbitrage under Basel I, CP-2 mandated the deduction from capital of retained first-loss positions and similar credit enhancements. The capital required for other securitized holdings de- pended on whether the bank had originated the securities or was an in- vestor in securitized assets of another issuer. Here again, the committee contemplated standardized, F-IRB, and A-IRB approaches. However, its securitization proposals for the IRB approaches were explicitly tentative and not fully developed.
While the committee had acquiesced to much of the criticism of CP-1, it held fast to its objective of introducing a capital charge for operational risk. Once again, though, CP-2 was more suggestive than fully elaborated.
The committee proposed to parallel its three approaches to credit risk capital requirements with three approaches to operational risk. The basic indicator approach would link the charge to a “single indicator that serves as a proxy for the bank’s overall risk exposure” (Basel Committee 2001a, paragraph 163). The committee suggested, but did not specify clearly, that gross income would be the relevant indicator. The standardized approach would also base the charge on an indicator that proxied for risk exposure, but here both the indicator and the percentage taken would vary by busi- ness line (e.g., corporate finance, retail banking). Finally, the internal meas- urement approach would permit a bank to use its own inputs to quantify the risk of operational losses in each of its business lines; the output would then be adjusted by a factor designated by the committee to yield the operational risk charge.
Notwithstanding the continued insistence of the committee that all three pillars were to be important regulatory tools, the disproportionate emphasis on pillar 1 had, if anything, increased in CP-2. The supervisory review process of pillar 2 was elaborated modestly from the CP-1 baseline through the articulation of four supervisory principles (box 4.3). The com- mittee also issued guidelines for supervisory review of the minimum standards for IRB eligibility, supervisory transparency and accountabil- ity, and the review of interest rate risk. The last of these represented an- other accommodation by the Basel Committee to criticism from banks, which had argued in reaction to CP-1 that the variety of their methods for
managing interest rate risk made it inappropriate for pillar 1 coverage.
The implications of some of these guidelines were potentially far-reaching.
For example, the CP-1 idea that banks should generally hold capital above the minimum required level was repeated. Another principle, seemingly echoing the US system of prompt corrective action, suggested that super- visors should intervene “at an early stage” to prevent capital from falling below regulatory minimum levels. Yet all of these guidelines seemed hor- tatory in nature and were generally addressed to supervisors, rather than banks.
The pillar 3 proposals in CP-2, on the other hand, built on an earlier committee paper to present a fairly detailed list of items to be disclosed by banks. The committee specified disclosures that it believed would facilitate the exercise of market discipline on risk-taking by banks. These require- ments, divided into core and supplementary disclosures, were grouped into 10 topical areas (box 4.4). Substantially greater disclosure was recom- mended for qualifying banks that elected an IRB approach. Despite the detail it provided, the committee’s expectations for implementation of pillar 3 were somewhat ambiguous. In particular, it noted that the legal authority of supervisors to require public disclosures by banks varied among member countries. Thus, although “some kind of enforcement response” was expected, that response might be indirect in some cases.
Box 4.3 Key principles of supervisory review in the second consultative paper
Principle 1: Banks should have a process for assessing their overall capital in relation to their risk profile and a strategy for maintaining their capital levels.
Principle 2: Supervisors should review and evaluate banks’ internal capital ade- quacy assessments and strategies, as well as their ability to monitor and ensure their compliance with regulatory capital ratios. Supervisors should take appro- priate supervisory action if they are not satisfied with the results of this process.
Principle 3: Supervisors should expect banks to operate above the minimum regulatory capital ratios and should have the ability to require banks to hold capital in excess of the minimum.
Principle 4: Supervisors should seek to intervene at an early stage to prevent capital from falling below the minimum levels required to support the risk char- acteristics of a particular bank and should require rapid remedial action if capital is not maintained or restored.
Source:Basel Committee (2001b).
The committee asked for comments on CP-2 by May 31, 2001. More than 250 comments were received in this four-month period.37Even be- fore that date, interested parties had spoken publicly, and privately to Basel Committee members, of their concerns. Many of the comments came from banks and associations representing banks or risk management officers, but comments were also received from other financial institutions, central banks and banking regulators around the world, international
Box 4.4 Areas for public disclosure by banks in the second consultative paper
䡲 Scope of application of capital accord within a banking group.
䡲 Breakdown of bank’s capital and supporting accounting procedures.
䡲 Breakdown of credit risk exposures and description of credit risk manage- ment techniques.
䡲 Identification and use of external credit assessment institutions (for banks using standardized approach).
䡲 Internal ratings–based (IRB) methodology, coverage, and ex post perform- ance of IRB methods, broken down for each portfolio (for banks using an IRB approach).
䡲 Exposures covered by credit risk mitigation measures, process for managing collateral, and methods for assuring continued creditworthiness of guaran- tors and credit derivative counterparties.
䡲 Methodology for calculating market risk capital requirements and, for banks using internal models approach, information about model, value-at-risk data, and backtesting.
䡲 Operational risk methodology and exposure(s).
䡲 Interest rate risk management technique and impact on earning and regula- tory capital of rate shocks.
䡲 Regulatory capital requirements for credit (including off-balance-sheet expo- sures), market, and operational risk; analysis of factors affecting capital ade- quacy; economic capital allocated to bank’s transactions, products, cus- tomers, business lines, or organizational units, as appropriate.
Source:Pillar 3 (Market Discipline), supporting document appended to Basel Committee (2001b).
37. Comments to CP-2 and CP-3 were all uploaded onto the Basel Committee website. Com- ments on CP-2 may be found at www.bis.org/bcbs/cacomments.htm (accessed May 15, 2008).
institutions, academics, think tanks, and credit rating agencies.38While it is obviously impossible to distill thousands of pages of comments into a fully representative synopsis of a few paragraphs, the most frequently re- curring themes can be summarized as follows.
First, commenters speaking directly or indirectly on behalf of banks generally applauded the committee’s embrace of IRB capital regulation, though quite a few continued to advocate faster movement toward the full use of internal models. Second, many academic or quasi-academic commenters expressed skepticism about the feasibility of the IRB approach to capital regulation, though the degree of skepticism varied considerably within this group. Third, while banks and others offered suggestions on a wide range of technical issues, a few provisions elicited criticism from a particularly high proportion of commenters. Among these were the pro- posal for a formulaic approach to a quantified capital charge for opera- tional risk and the proposal to require credit risk capital charges for ex- pected, as well as unexpected, losses.39Fourth, some commenters took issue with the risk-weighting formulas under the IRB approaches (and, in some cases, the standardized approach). Most frequently mentioned were the effects on firms in emerging markets and on small and medium-sized enterprises in the Basel Committee countries themselves. The committee’s use of a “scaling factor” that multiplied by a factor of about 1.5 the proba- bility of default value generated by the bank’s internal system was also criticized. Fifth, many commenters noted that the CP-2 proposals were incomplete in important respects and thus they could not judge the com- mittee’s proposals as a whole. Sixth, many banks complained that the menu of disclosures proposed under pillar 3 would require banks to re- veal much proprietary information and, in any case, would be needlessly expensive for the likely value to investors.
Seventh, and ultimately most important, many banks complained that regulatory capital would rise under the CP-2 proposals. The banks’ con- clusion, based on some quick number crunching, was borne out by the committee’s second quantitative impact study (QIS). QIS-2 had been initi- ated in April 2001, a few months after release of the CP-2. The results, re- ported in November of that year (Basel Committee 2001c), revealed that regulatory capital levels would increase under each of the three CP-2 ap-
38. The very fact that certain entities submitted comments raised interesting questions. For example, the Federal Reserve Banks of Chicago and Richmond both sent letters to the Basel Committee, giving rise to the inference that the Board of Governors and the New York Fed had not fully involved the rest of the Federal Reserve System in developing their Basel Com- mittee positions.
39. This proposal was particularly unwelcome to banks in the United States and other coun- tries where the practice is to set aside reserves for expected losses. The banks argued that the CP-2 proposal would, in effect, require them to provision for expected losses twice—once in their reserves and once in their capital requirements.
proaches (see table 4.2 for a summary of the results).40For the A-IRB ap- proach, the capital charge for credit risk alone would decline modestly, but the proposed new requirement for operational risk would result in in- creased overall regulatory capital. The overall percentage increases for the standardized approach were in the mid-teens. Most embarrassing for the Basel Committee, the overall percentage increases for the foundational IRB approach were even higher. Thus, under the CP-2 proposal, banks would have no incentive to shift from a standardized to the F-IRB approach.
As noted earlier, the committee’s stated intentions were to maintain average regulatory capital levels unchanged from Basel I under the new standardized approach and to bring about a modest reduction from Basel I levels for banks adopting one of the IRB approaches. However, in light of the implications of the CP-2 methodologies for minimum capital levels, the committee now had a credibility problem. It did not seem to understand the effects of its own proposals. Far from being the “dream”
for banks that Bill McDonough had promised, some banking executives, in the words of one, saw “themselves being railroaded into the new ap- proach at a lot of cost.”41
Table 4.2 Results of second quantitative impact study change in capital requirements under second consultative paper proposals (percent)
Standardized Foundation IRB Advanced IRB Group Credit Overall Credit Overall Credit Overall G-10
Group 1 6 18 14 24 ⫺5 5
Group 2 1 13
European Union
Group 1 6 18 10 20 ⫺1 9
Group 2 ⫺1 11
Other (non-G-10, non-EU) 5 17 IRB ⫽internal ratings–based
Notes: The second quantitative impact study (QIS-2) exercise involved 138 banks from 25 coun- tries. Group 1 banks were diversified, internationally active banks. Group 2 banks were smaller or more specialized banks. Banks from Basel Committee member countries that were also European Union (EU) members were included in both the G-10 and EU categories in the reported results.
Source:Basel Committee (2001c).
40. By the time the Basel Committee released the QIS-2 results, it had already decided to re- duce the operational risk charge significantly. Thus, when the banks initially ran their num- bers after release of the CP-2 proposals, they would have observed a substantially greater in- crease in capital requirements.
41. See John Willman, “Industry Backs Postponing New Rules on Amount of Capital Banks Must Hold,” Financial Times, June 25, 2001.